JP2002024258A - Image acquisition device and method, and computer- readable recording medium recorded with image acquisition processing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain image data of plural documents by the documents without using any document sectioning original when the unspecified documents are read by an image scanner having an automatic document feeding function. SOLUTION: Image data are generated by reading the documents by an image input part 10 and stored in an image storage part 20. Then a feature quantity calculation part 30 calculates a feature quantity needed to decide document sectioning is calculated from all the image data stored in the image storage part 20. A document sectioning decision part 40 decides the document sectioning according to the feature quantity calculated by the calculation part 30. When the document sectioning result displayed at a display part 80 is different from a desired result, the document sectioning is corrected through an input part 70.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image scanner having an automatic paper feed function, which reads a plurality of documents by using a document delimiter. The present invention relates to an image acquisition apparatus and an image acquisition method to acquire an image, and a computer-readable recording medium recording an image acquisition processing program.

[0002]

2. Description of the Related Art In recent years, when a document composed of one or more paper originals is efficiently digitized over a plurality of copies, a process of continuously reading the originals using an image scanner having an automatic paper feed function is widely performed. Have been done. At this time, in order to manage the read document image data for each document, it is necessary to detect between documents by using some means. Several approaches have been proposed to meet such needs.

For example, in Japanese Patent Laid-Open Publication No. 1-162474, a specific document delimiting document is inserted between documents before reading a plurality of document documents, so that the document can be read continuously by an image scanner. In this case, an electronic file device that recognizes the document for document separation and registers the image of the document for each document is shown.

In Japanese Patent Application Laid-Open No. 7-287747, a document delimiter mark is added to the back of a document serving as a document delimiter in advance, so that both sides of the document are read by an image scanner and the document delimiter mark is recognized. 1 shows an electronic file device for registering a document.

As another approach, Japanese Patent Laid-Open No.
Japanese Patent Application Laid-Open No. 21380 discloses an electronic file device that cuts out a previously set character recognition area from image data read by an image scanner, performs character recognition, and determines a document break based on the character recognition result.

[0006]

However, the method of determining a document break using the document for document break has a great advantage that it does not depend on the type of the document to be read. However, there is a problem in that a document for document separation corresponding to the number of documents must be prepared and inserted between the documents, and the user's work load increases as the number of documents increases.

The method of adding a document break mark to the back side of a document serving as a document break has the advantage that there is no need to insert a document break document, but there is a limitation that a document to be read cannot be rewritten. Originally, when information is described on both sides of the original, the mark cannot be added, and there is a problem that the mark cannot be used for a document composed of such an original.

Further, in the method using character recognition, a title, a total number of pages, and the like are set at predetermined positions of a document.
There must be a character string that satisfies the conditions that can be used when determining document breaks, and it is not a method that can be used for arbitrary manuscripts. There is a problem that it is difficult to use a document.

The present invention has been made in view of the above circumstances, and when reading an unspecified plurality of documents by an image scanner having an automatic paper feed function, the images of the plurality of documents are used without using a document for document separation. It is an object of the present invention to provide an image acquisition device and an image acquisition method capable of acquiring data in document units, and a computer-readable recording medium on which an image acquisition processing program is recorded.

[0010]

An image processing apparatus according to the present invention for achieving the above object has the following features.
That is, the invention according to claim 1 of the present application is a document image reading means for sequentially reading documents of a plurality of documents, a feature value calculating means for calculating a feature value of a document image read by the document image reading means, and a feature value calculating means Document separation determining means for determining the document break of the document image read by the document image reading means based on the characteristic amount of the document image calculated by the above. This makes it possible to determine a document break without inserting a document break document between documents in advance or modifying a document serving as a document break, thereby greatly reducing the burden imposed on the user. .

According to a second aspect of the present invention, there is provided a document image reading means for sequentially reading a plurality of documents, a document number input means for inputting the number of documents to be read by the document image reading means, and a document image reading means. A feature amount calculating unit for calculating the feature amount of the document image obtained by reading the document image, and a document image reading based on the number of documents input by the document number input unit and the feature amount of the document image calculated by the feature amount calculating unit. Means for determining a document break of a document image read by the means. Thereby, said claim 1
In addition to the advantages of the invention according to the first aspect, the number of documents is known in advance, so that a more accurate document segmentation determination can be performed.

The invention according to claim 3 of the present application calculates a plurality of feature amounts by the feature amount calculating means. As a result, even when a document of an unspecified type is read and the document segmentation is determined, the document segmentation determination with higher accuracy can be performed as compared with the case where a single feature amount is used.

In the invention according to claim 4 of the present application, the document segment determination means selects a feature value suitable for dividing a document from a plurality of feature values calculated from the document image by the feature value calculation device, This is to determine a document break.
This eliminates unnecessary features for document separation determination depending on the type of document, not only speeding up the document separation determination process itself, but also reduces the adverse effects of unnecessary features on document separation determination. Can be avoided.

Further, the invention according to claim 5 of the present application further comprises a document type registration unit for registering a feature amount selected by the document break determination unit and suitable for separating a document in association with an arbitrary document type name. Have With this, if a document of the type that has undergone the document break determination process is used, when the same type of document is read later, it is possible to always select an optimal feature amount for the document break determination simply by specifying the registered document type name. Thus, the processing time can be greatly reduced as compared with the method of selecting the characteristic amount each time. In addition, external factors when reading the original, such as when the original is partially dirty, have changed. Can be.

[0015] In the invention according to claim 6 of the present application, the document segmentation judging means performs weighting suitable for segmenting the document, on the feature quantity of the number of characters calculated from each document image by the feature quantity calculating means. After that, the break of the document is determined. As a result, even when the types of documents are various, the integrated determination can be performed by using each characteristic amount exquisitely, so that a more accurate document segmentation determination can be performed.

The invention according to claim 7 of the present application provides a document type for registering a weight set for each feature amount most suitable for separating a document, set by a document separation determining unit, in association with an arbitrary document type name. It further has registration means.
As a result, if a document of the type that has been subjected to the document break determination process once, it will always be the best for the document break determination for each feature amount by simply specifying the registered document type name when reading the same type of document later. Weighting can be performed, and an effect of shortening the processing time and an effect of obtaining a stable document break determination can be expected.

The invention according to claim 8 of the present application provides a document break correction in which a document break determination result determined by the document break determination means is displayed and a user corrects the document break position in accordance with the displayed determination result. Means. This makes it possible to easily correct the document break even if the result of the document break determination is different from the desired result.

The invention according to a ninth aspect of the present invention provides a document break correction in which a result of a document break determined by the document break determining means is displayed, and a user corrects a document break position in accordance with the displayed determination result. Further comprising means,
In this document break correction means, when the user makes a correction to some of the document breaks, the user selects a feature amount suitable for calculating the corrected document break, and according to the selected feature amount. , The remaining document breaks are determined again. As a result, even if the document break determination result needs to be corrected, the user only needs to correct the document break at one place, and when the correction is reflected, the effective feature amount is back calculated and based on the back calculated feature amount. Therefore, a document break determined to require another correction can be automatically corrected, so that the burden imposed on the user can be greatly reduced.

The invention according to claim 10 of the present application provides a document segmentation correction which displays a document segmentation determination result determined by the document segmentation determination means and corrects a document segmentation position by a user according to the displayed determination result. Further comprising means,
In this document break correction means, when the user makes a correction to a part of the document break, a weighting coefficient for each feature amount suitable for calculating the corrected document break is calculated. The remaining document breaks are re-determined using the feature weighted by the calculated coefficients. As a result, even if it is necessary to correct the document break determination result, the user only needs to correct the document break at one place, and the weighting coefficient for each feature amount effective in reflecting the correction is calculated backward. Since the document break that is determined to require another correction can be automatically corrected based on the feature weighted by the weighted coefficient, the burden imposed on the user can be greatly reduced.

In the invention according to claim 11 of the present application, the document break correction means presents a plurality of document break candidates for one document break. Thus, even if the document break needs to be corrected, the user only has to select a correction candidate, so that the burden imposed on the user in the document break correction process can be greatly reduced.

Further, the invention according to claim 12 of the present application further comprises a document type input means for inputting the type of the document to be read, wherein the characteristic amount calculating means is adapted to respond to the type of the document input by the document type input means. Is used to calculate the feature amount.
As a result, in the case where the type of the document to be read is clear in advance, it is possible to always calculate only the characteristic amount necessary for the document segmentation determination of the type by simply designating the document type name. The calculation time can be saved.

The invention according to claim 13 of the present application further comprises a document type input means for inputting the type of the document to be read, and the document delimiter judging means determines the type of the document input by the document type input means. A feature amount is selected in accordance with this. In this way, when the type of a document to be read is clear in advance, not only is it possible to always select a feature amount suitable for determining the type of document delimiter simply by specifying the document type name, but also to select a feature amount. The time required can be saved. In addition, the external factors when reading the original, such as when the original is partially dirty, have fluctuated. Can be.

The invention according to claim 14 of the present application further comprises a document type input means for inputting the type of the document to be read, and the document delimiter judging means determines the type of the document input by the document type input means. The weighting for each feature amount is changed accordingly. In this way, when the type of the document to be read in advance is clear, not only is it possible not only to specify the document type name but always to optimally weight each feature amount for the document segmentation determination, Can be saved in calculating the weighting factor for.

Further, according to the invention according to claim 15 of the present invention, the document segmentation judging means sets a sample corresponding to the feature amount calculated from each image data in the pattern space constituted by the feature amount calculated by the feature amount calculating means. Clustering is performed, and the number of document breaks is set according to the number of obtained clusters. Thus, in a case where each cluster represents one document, highly accurate document segmentation determination can be performed.

In the invention according to claim 16 of the present application, in the pattern space constituted by the feature amount calculated by the feature amount calculating means, the document segmentation judging means extracts a sample corresponding to the feature amount calculated from each image data. Clustering is performed, and image data corresponding to a sample belonging to an arbitrary cluster is determined to be a document delimiter. As a result, when a certain cluster is made of an image that can be used as a document break, such as a cover image of a document, the document break can be determined using this cluster.

According to the seventeenth aspect of the present invention, the document break determination means arranges the characteristic values of the image data of the document calculated by the characteristic value calculation means in the order of reading the document by the document image reading means. When comparing the value of the feature value calculated from the image data of an arbitrary document with the value of the feature value calculated from the image data of at least one of the document read immediately before or the document read immediately after, If the difference is larger than a predetermined threshold, the image data is determined to be a document break. As a result, even when only one type of feature is used for the document break determination, the break of the document can be easily determined if the feature is suitable for determining the break of the document.

Further, the invention according to claim 18 of the present application is characterized in that the characteristic amount calculated by the characteristic amount calculating means is spatial frequency information of the document image. Thereby, for example, the screen ruling of the document can be used as the feature amount.

Further, the invention according to claim 19 of the present application is characterized in that the characteristic amount calculated by the characteristic amount calculating means is density information of a document image. Thus, for example, when the document is read in color, the color arrangement of the document can be used as the feature amount by determining the density value of the pixel of each of the RGB plane images.

Further, the invention according to claim 20 of the present application is characterized in that the characteristic amount calculated by the characteristic amount calculating means is the layout information of the document. As a result, for example, the arrangement information of characters and photographs obtained by analyzing the image data of the document, the combination direction of characters, the number of characters in one line / column, the size / type of characters, and the like can be used as the feature amounts.

According to a twenty-first aspect of the present invention, there is provided a document image reading means for sequentially reading documents of a plurality of documents, a feature value calculating means for calculating a feature value of a document image read by the document image reading means. A document segment determining unit that determines a document segment of the document image read by the document image reading unit based on the feature amount of the document image calculated by the feature value calculating unit; and a document obtained by the document image reading unit. Document storage means for dividing an image into each document according to the document break determined by the document break determination means, and storing a document image corresponding to each document as an individual electronic file. This makes it possible to determine a document break without inserting a document break manuscript between each document in advance or modifying a document serving as a document break, so that a plurality of documents can be accurately classified and classified. Each document can be stored as an electronic file.

Further, the invention according to claim 22 of the present application is based on the step of sequentially reading the originals of a plurality of documents, the step of calculating the characteristic amount of the original image obtained by reading, and the step of reading the calculated characteristic amount of the original image. And a step of determining a document break of a plurality of documents in a document image. This makes it possible to determine a document break without inserting a document break document between documents in advance or modifying a document serving as a document break, thereby greatly reducing the burden imposed on the user. .

Further, in the invention according to claim 23 of the present application, when reading a plurality of documents, a process of inputting the number of documents to be read and a process of sequentially reading documents of a plurality of documents are obtained by reading. An image acquisition method comprising: a step of calculating a feature amount of a document image; and a step of determining a document break of a plurality of documents in the document image based on the number of input documents and the calculated feature value of the document image. is there.
Thereby, in addition to the advantage of the invention according to claim 22,
Since the number of documents is known in advance, it is possible to perform more accurate document segmentation determination.

Further, the invention according to claim 24 of the present application is a step of sequentially reading a document of a plurality of documents, a step of calculating a characteristic amount of the read original image, and a step of calculating the characteristic amount of the calculated original image. Determining the document breaks of a plurality of documents in a document image, and separating the document images into individual documents according to the determined document breaks, and storing the document images corresponding to the respective documents as individual electronic files. An image acquisition method comprising: This makes it possible to determine a document break without inserting a document break manuscript between each document in advance or modifying a document serving as a document break, so that a plurality of documents can be accurately classified and classified. Each document can be stored as an electronic file.

According to a twenty-fifth aspect of the present invention, there is provided a procedure for sequentially reading manuscripts of a plurality of documents, a procedure for calculating the characteristic amount of the read manuscript image, and a method for calculating the characteristic amount of the manuscript image. And a computer-readable recording medium that records an image acquisition processing program for causing a computer to execute a procedure for determining document breaks of a plurality of documents in a document image. By executing these procedures, document breaks can be determined without inserting a document break document between documents in advance or modifying the document serving as a document break, greatly reducing the burden on the user. can do.

In the invention according to claim 26 of the present application, when reading a plurality of documents, a procedure for inputting the number of documents to be read, a procedure for sequentially reading the documents for a plurality of documents, and a procedure for reading the plurality of documents are provided. Image acquisition for causing a computer to execute a procedure for calculating a feature amount of a document image and a procedure for determining a document break of a plurality of documents in the document image based on the number of input documents and the calculated feature value of the document image It is a computer-readable recording medium on which a processing program is recorded. By executing these procedures, in addition to the advantage of the invention according to claim 25, since the number of documents is known in advance, it is possible to perform more accurate document segmentation determination.

The invention according to claim 27 of the present application provides a procedure for sequentially reading manuscripts of a plurality of documents, a procedure for calculating the characteristic amount of the read manuscript image, and a method for calculating the characteristic amount of the manuscript image. A procedure for determining document breaks of a plurality of documents in a document image, a procedure for separating the document images into respective documents according to the determined document breaks, and storing the document images corresponding to each document as individual electronic files And a computer-readable recording medium in which an image acquisition processing program for causing a computer to execute the above is recorded. By performing these procedures, document breaks can be determined without inserting a document break manuscript between each document in advance or modifying the manuscript serving as a document break, so that multiple documents can be accurately classified. The document can be stored as an electronic file for each of the classified documents.

[0037]

Embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a schematic configuration diagram illustrating an image acquisition device according to an embodiment of the present invention. That is, the image acquisition device according to the present embodiment includes the image input unit 1
0, an image storage unit 20, a feature amount calculation unit 30, a document break determination unit 40, a document storage unit 50, a document break correction unit 60, an input unit 70, a display unit 80, a control unit 90 for controlling the operation of the entire apparatus, And a bus 100 connecting the components.

The image input unit 10 includes an image scanner having an automatic paper feed function, reads a plurality of document originals,
Input as image data (original image). The image scanner of this apparatus is capable of reading both sides of a document and a color document. It should be noted that the image scanner may be connected exclusively to this apparatus, or an image scanner operating as a network scanner may be used as one user.

The image storage section 20 sequentially stores the image data of the document input by the image input section 10. An image memory for storing image data exclusively, a RAM or a hard disk used for performing various arithmetic processing,
A general-purpose recording medium such as a removable disk may be used.

The feature amount calculation unit 30 calculates a feature amount necessary for document segmentation determination from image data being input by the image input unit 10 or image data stored in the image storage unit 20 in advance.

The document break determination unit 40 includes the feature amount calculation unit 3
At 0, image data serving as a document break is determined based on the feature amount calculated from the image data.

The document storage section 50 stores the image data stored in the image storage section 20 as electronic data in document units according to the document break determination by the document break determination section 40. The document storage unit 50 may be an image memory for storing image data exclusively, as in the case of the image storage unit 20, or a RAM used for performing various arithmetic processing.
It may be a general-purpose recording medium such as a hard disk, a removable disk, or the like. Alternatively, only the information for enabling the image data stored in the image storage unit 20 to be handled in document units without storing the image data itself may be stored.

The document break correction unit 60 corrects the document break determined by the document break determination unit 40 when it is not the one desired by the user. More specifically,
First, the result of the document break determination by the document break determination unit 40 is displayed to the user through the display unit 80, and the user is asked whether correction of the document break is necessary. The user operates the display unit 80
With reference to the document segmentation determination result displayed in step (1), an appropriate instruction is given to the document segmentation modification unit 60 through the input unit 70.

The input unit 70 receives a series of instruction operations from a keyboard, a mouse, and the like by the user of the apparatus, and the display unit 80 displays necessary information to the user of the apparatus each time. . That is, the input unit 70
And the display unit 80 are the core of the interface between the apparatus and the user.

The control section 90 controls the entire apparatus according to the present embodiment.

Next, the operation (image acquiring method) of the entire image acquiring apparatus shown in FIG. 1 will be described with reference to the flowchart shown in FIG.

First, in step S101, it is determined whether the number of documents to be read has been designated by the user.
Here, when the number of read documents is designated by the user, the value is held (step S102).

The stored number of documents will be used later for determining a document break. If the number of documents to be read is not specified by the user, the number of documents is determined at the same time as the document break is determined in the later document break determination processing (step S103).

Subsequently, in step S104, it is determined whether or not the user has instructed the feature amount of the image data to be used in the document break determination processing. If there is an instruction from the user, the fact is set as a condition for the document break determination process (step S105).

As a method of designating the characteristic amount, a method that can be quantitatively expressed, for example, a hue and a screen ruling used in image data, and a qualitative method such as a layout of a document and a combination direction of characters are selected. Alternatively, weighting may be used. Instead of directly specifying these feature amounts, the types of frequently used documents may be registered in advance, and the method of using the feature amounts may be changed according to the types of these documents.

FIG. 3 is a diagram showing an example of prompting the user through the display unit 80 as an example of an instruction relating to the feature amount of image data used in the document break determination processing. 3A and 3B, it is assumed that the document to be read by the apparatus is a standard document or an atypical document, and is used in a case where the document can be specified in advance. By doing so, the method of using the feature amount is indirectly determined.

FIG. 3A shows an example in which a type of document whose cover is almost fixed is selected, and FIG. 3B shows an example in which a non-standard type of document such as a product catalog is selected. I have. On the other hand, FIG. 3C shows an example in which items such as the hue and the layout structure of the image data are directly designated as the feature amounts used for determining the document break.

If there is no instruction from the user for the feature value used in the document break determination process in step S104, the document break is determined in an integrated manner using a predetermined feature value (step S106).

Next, in step S107, the image input unit 10 reads a plurality of documents and generates image data. At this time, two pieces of image data are generated when both sides are read for one document, and one image data is generated when only one side is read. Then, the generated image data is stored in the image storage unit 20 in the order in which the originals are read.

[0055] Incidentally, the read or only one or single-sided reading both sides of a document when reading the document, the user may specify each time, whereas previously either as a default value when the document reading by the device May be set.

Next, in step S108, a feature amount is calculated from all the image data input in step S107. The feature amount calculation unit 30 sequentially acquires the images stored in the image storage unit 20 and calculates the feature amount. If it takes a relatively long time to calculate the feature amount, the display unit 80 may display the detailed progress of the feature amount calculation process. Note that the feature amounts calculated from each image are derived in steps S105 and S106.

Next, in step S109, based on the characteristic amount calculated in step S108, step S107
Calculate the breaks of multiple documents input in. At this time,
If the number of documents input in step S102 is specified by the user, the number of documents is used to determine a document break.

If a relatively long time is required for the document segmentation determination process, similar to the feature amount calculation process in step 8108, a detailed progress of the document segmentation determination process is displayed on the display unit 80. Is also good.

The calculated document break is displayed on the display unit 80.
Notify the user through. Details of the document break determination processing in step S109 will be described later.

In step S110, the user ← the user
It is determined whether or not the result of the document segmentation determination in step S109 displayed as 0 is the desired one, and the result is transmitted to the apparatus through the input unit 70.

If it is determined that the result of the determination of the document break in step S109 is not the desired one and needs to be corrected, the process of correcting the document break is performed in step S111.

Here, the user may manually correct all document breaks, or if the user makes corrections to the first document break, a feature most suitable for calculating the corrected document break. The amount and the weighting coefficient for each feature amount may be calculated backward, and the remaining document breaks may be re-determined based on the result without bothering the user.

When the determination of the document break is completed for all the input document documents, the document image data is stored in document units (step S112). As described above, without storing the image data itself, only the information for enabling the image data stored in the image storage unit 20 to be handled in document units may be stored.

When all the desired document breaks have been obtained, a series of feature amounts used to calculate each document break and a coefficient for weighting the feature amounts can be reused in step S105. May be added with an arbitrary document type name and recorded.

Next, the details of the operation for determining a document break shown in step S109 will be described with reference to the flowchart of FIG. First, in step S201, it is determined whether the number of documents to be read has been designated by the user as in step S101.

If the number of documents to be read is specified by the user, the maximum number of clusters in clustering, which will be performed later in the pattern space composed of the feature amounts calculated in step S108, is set to the number of read documents. (Step S202).

If it is determined in step S201 that the number of documents to be read has not been specified by the user, then in step 8203, as in step S104, the user has instructed the feature amount of the image data to be used in the document break determination processing. Determine whether

In step S203, when the user instructs the feature amount of the image data to be used in the document break determination process, the maximum number of clusters is set to a predetermined value based on the instruction and the number of the originals actually read (step S203). Step S204). Conversely, if no instruction has been given from the user, the maximum cluster number is set to the number of documents actually read (step S205).

Next, when the setting for weighting each feature is selected in step S105, weighting is performed on each feature calculated in step S206.

Next, in step S207, clustering is performed on a sample group having the value of the feature calculated from the image data of each document in the pattern space around each feature. At this time, the maximum number of clusters obtained as a result is determined in steps S202, S204,
Alternatively, it does not exceed the maximum value set in step S205.

In step S208, step S207
It is determined whether or not the result of the clustering performed in is good.

FIG. 5 is a diagram showing an example of sample distribution in a pattern space having two feature amounts (X ₁ , X ₂ ) as axes, in order to easily explain the quality judgment of the clustering result. .

FIG. 5A shows an example in which good clustering can be performed. In this example, two clusters ω ₁ and ω ₂ are obtained by clustering. On the other hand, FIG. 5B shows an example in which it is difficult to perform appropriate clustering for the samples in FIG. 5 in the pattern space including the feature amounts of X ₁ and X ₂ .

Next, if it is determined in step S208 that the clustering is good, it is determined in step S209 whether or not any of the obtained clusters includes a cluster composed of image data samples that can serve as document delimiters. judge.

For example, the cluster ω ₁ in FIG.
Almost original is continuously reading order, if the sample belonging to the cluster omega ₂ is such that it is discontinuous in the reading order of the document in the cluster, the sample, for example a plurality of documents belonging to the cluster omega ₂ samples in the cluster belonging to There is a possibility that the document corresponds to a document break that appears once on several pages, such as a cover document image when the document is read.

Here, it is determined that the sample belonging to the cluster ω ₂ represents such a document break (step S210). Thus, there are three documents in the example of FIG.

On the other hand, when the samples belonging to the clusters ω ₁ and ω ₂ in FIG. 5A are continuous in the reading order of the original in each cluster, one of the clusters indicates a document break. Cannot be determined. In such a case, it is more appropriate to consider that the two clusters ω ₁ and ω ₂ each represent a document.

The number of obtained clusters is determined in step S
Even when the number of documents coincides with the number set in step 102, it may be more appropriate to consider that each cluster itself represents an individual document as described above.

Here, as a document serving as a document break,
Image data corresponding to the first sample or the last sample read from each cluster is set as a document break (step S21).
1). As a result, two documents exist in the example of FIG.

On the other hand, as shown in FIG. 5B, if it is determined in step S208 that the clustering is difficult or the clustering result is not good, the re-clustering process can be performed by changing the weight for each feature amount. It is determined whether or not it is (step S212).

If it is determined that the re-clustering process can be performed by changing the weight for each feature, the process returns to step S206 to change the weight for each feature. Is performed again. This determination and the change of the weighting for each feature amount may be manually performed by the user through the input unit 70, or may be automated according to a preset procedure.

In step S212, it is determined that it is difficult to change the weight for each feature and perform the re-clustering process.
, One document is set as one document, or all read documents are set as one document. These may be designated each time through the input unit 70, or one of them may be set as an initial value when such a situation occurs.

In FIG. 5, the clock in the pattern space is
As an example of rastering, the feature quantity X ₁And X_TwoUsing
As described above, the method of specifying the feature amount in step S105
Depending on the method, more features may be used,
The amount may be one.

When there is only one type of feature, the document break is determined relatively easily from the relationship between the feature and the reading order of the original in step S107, as shown in FIG. 6, without performing clustering processing in the pattern space. it can.

For example, it is now assumed that the feature quantity X ₁ in FIG. 6 is obtained by digitizing the layout structure of image data according to a certain rule. At this time, FIG. 6A shows the pages represented by the samples S1 and S7 as a cover or a page similar thereto, and the other samples are two documents each representing a page having substantially the same layout structure. This is an example that exists.

FIG. 6B shows an example in which documents having two different layouts represented by samples S1 to S8 and samples S9 to S7 exist. As described above, in both FIGS. 6A and 6B, the feature amount X _{1 is used.}
By setting the threshold value TH in, the document serving as a document break can be determined.

In the present invention, the above operation may be recorded on a computer-readable recording medium (for example, a CD-ROM) as an image acquisition processing program.

[0088]

As described above, according to the present invention, when a plurality of unspecified documents are read by an image scanner having an automatic sheet feeding function, the image data of the plurality of documents can be read without using a document separating document. It can be obtained for each document. As a result, it is possible to greatly reduce the burden on the user that occurs when converting a large number of documents into dragons. Also, since the feature amount for determining the document break and the weight for the feature amount are changed according to the type of the document,
Various types of documents can be handled without adding a predetermined symbol or character string such as a document delimiter mark to a document in advance.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram illustrating an image processing apparatus according to an embodiment.

FIG. 2 is a flowchart illustrating an operation of the entire image processing apparatus.

FIG. 3 is a diagram showing a display example when a feature amount is instructed.

FIG. 4 is a flowchart illustrating an operation of document break determination.

FIG. 5 is a diagram illustrating clustering using a plurality of feature amounts when performing document segmentation determination.

FIG. 6 is a diagram illustrating threshold processing using a single feature amount when performing document segmentation determination.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 10 ... Image input part, 20 ... Image storage part, 30 ... Feature amount calculation part, 40 ... Document separation determination part, 50 ... Document storage part, 6
0: Document break correction unit, 70: Input unit, 80: Display unit,
90: control unit, 100: bus

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G06T 7/00 250 G06T 7/00 250 H04N 1/21 H04N 1/21 F-term (Reference) 5B009 SA00 5B050 BA10 BA16 EA04 EA09 FA02 FA13 GA08 5B075 ND07 NR12 PP02 PP03 PP04 PQ02 UU06 5C073 AA06 AB02 5L096 AA02 BA18 CA21 FA81 GA51 JA11 MA07

Claims (27)

[Claims] 1. A document image reading means for sequentially reading documents of a plurality of documents; a feature value calculating means for calculating a feature value of a document image read by the document image reading means; An image acquisition device comprising: a document segment determining unit configured to determine a document segment of a document image read by the document image reading unit based on the feature amount of the document image. 2. A document image reading means for sequentially reading a plurality of documents, a document number input means for inputting the number of documents to be read by the document image reading means, and a document image obtained by reading the document image reading means. A feature amount calculating unit that calculates the feature amount of the document, and the document image reading unit reads the document number based on the number of documents input by the document number input unit and the feature amount of the document image calculated by the feature amount calculating unit. An image acquisition apparatus comprising: a document break determining unit that determines a document break of an obtained document image. 3. The image acquisition apparatus according to claim 1, wherein said characteristic amount calculating means calculates a plurality of characteristic amounts. 4. The method according to claim 1, wherein the document segment determining unit selects a feature amount suitable for dividing the document from a plurality of feature amounts calculated from the document image by the feature amount calculating unit to determine a document segment. The image acquisition device according to claim 3, wherein: 5. A document type registering means for registering a feature amount selected by the document segmentation determining means and suitable for separating a document in association with an arbitrary document type name. The image acquisition device according to the above. 6. The document delimiter determining means determines the document delimiter after weighting each of the plurality of feature values calculated from each document image by the feature value calculating means suitable for separating the document. 4. The method according to claim 3, wherein
The image acquisition device according to the above. 7. A document type registering means for registering the weighting for each feature amount most suitable for separating a document set by the document segmentation determining means in association with an arbitrary document type name. The image acquisition device according to claim 6. 8. A document delimiter for displaying a document delimiter determination result determined by the document delimiter determiner and correcting a document break position by a user according to the displayed determination result. The image acquisition device according to claim 1. 9. A document delimiter that displays a result of document delimitation determined by the document delimiter and corrects a document break position by a user according to the displayed determination result. The correction means selects a feature amount suitable for calculating the corrected document break when the user corrects a part of the document break, and according to the selected feature amount, 4. The image acquisition apparatus according to claim 3, wherein a document break is determined again. 10. A document delimiter that displays a document delimiter determination result determined by the document delimiter determiner and corrects a document break position by a user according to the displayed determination result. The correction unit calculates a weighting coefficient for each feature amount suitable for calculating the corrected document break when the user makes a correction to a part of the document break, and calculates the weighted coefficient by the calculated coefficient. 4. The image acquisition apparatus according to claim 3, wherein the remaining document breaks are re-determined using the weighted feature amount. 11. The image according to claim 8, wherein the document break correction unit presents a plurality of document break candidates for one document break. Acquisition device. 12. A document type input unit for inputting a type of a document to be read, wherein the characteristic amount calculating unit calculates a characteristic amount according to the type of the document input by the document type input unit. The image acquisition device according to claim 1 or 2, wherein 13. A document type input unit for inputting a type of a document to be read, wherein the document break determination unit selects a feature amount according to the type of the document input by the document type input unit. The image acquisition device according to claim 3, wherein 14. A document type input unit for inputting a type of a document to be read, wherein the document break determination unit changes a weight for each feature amount according to the type of the document input by the document type input unit. The image acquisition device according to claim 3, wherein: 15. The document break determination unit clusters a sample corresponding to a feature amount calculated from each image data in a pattern space formed by the feature amount calculated by the feature amount calculation unit, and obtains a cluster. 3. The image acquisition apparatus according to claim 1, wherein the number of document breaks is set in accordance with the number of documents. 16. The document break determination unit clusters a sample corresponding to a feature amount calculated from each image data in a pattern space configured by the feature amount calculated by the feature amount calculation unit, and forms a cluster into an arbitrary cluster. 3. The image acquisition apparatus according to claim 1, wherein the image data corresponding to the sample to which the image data belongs is determined to be a document delimiter. 17. The document separation determining unit calculates, when the values of the feature amounts calculated by the feature amount calculating unit are arranged in the order in which the document is read by the document image reading unit, from image data of an arbitrary document. The value of the feature value is compared with the value of the feature value calculated from the image data of at least one of the document read immediately before or the document read immediately after, and the difference is larger than a predetermined threshold value. 3. The image acquisition device according to claim 1, wherein in the case, the image data is determined to be a document delimiter. 18. The image acquisition apparatus according to claim 1, wherein the feature amount calculated by the feature amount calculation unit is spatial frequency information of a document image. 19. The apparatus according to claim 1, wherein the characteristic amount calculated by the characteristic amount calculating unit is density information of a document image. 20. The apparatus according to claim 1, wherein the feature amount calculated by the feature amount calculation unit is layout information of a document. 21. A document image reading means for sequentially reading documents of a plurality of documents, a feature value calculating means for calculating a feature value of a document image read by the document image reading means, and a calculation by the feature value calculating means. Document delimiter determining means for determining a document break of a document image read by the document image reading means based on the feature amount of the read document image; and a document image obtained by reading the document image by the document image reading means. An image acquisition apparatus, comprising: a document storage unit that divides each document according to the document break determined by the break determination unit and stores a document image corresponding to each document as an individual electronic file. 22. A process for sequentially reading the originals of a plurality of documents; a process for calculating the characteristic amount of the read original image; and a process for calculating the characteristic amount of the original image based on the calculated characteristic amount of the original image. Determining a document break. 23. In reading a plurality of documents, a process of inputting the number of documents to be read, a process of sequentially reading the documents of the plurality of documents, and calculating a characteristic amount of a document image obtained by reading. An image acquisition method, comprising: determining a document break of a plurality of documents in the document image based on the input number of documents and the calculated feature amount of the document image. 24. A process for sequentially reading documents of a plurality of documents, a process for calculating feature values of the read document images, a plurality of documents in the document images based on the calculated feature values of the document images. And a step of dividing the document image for each document according to the determined document break and storing the document image corresponding to each document as an individual electronic file. Characteristic image acquisition method. 25. A procedure for sequentially reading manuscripts for a plurality of documents, a procedure for calculating feature quantities of the read manuscript images, and A computer-readable recording medium in which an image acquisition processing program for causing a computer to execute a procedure for determining a document break is recorded. 26. In reading a plurality of documents, a procedure for inputting the number of documents to be read, a procedure for sequentially reading the plurality of documents, and calculating a characteristic amount of the document image obtained by reading. An image acquisition processing program for causing a computer to execute a procedure and a procedure of determining a document break of a plurality of documents in the document image based on the input number of documents and the calculated feature amount of the document image. Computer readable recording medium. 27. A procedure for sequentially reading manuscripts of a plurality of documents, a procedure for calculating a feature amount of the manuscript image obtained by reading, a plurality of documents in the manuscript image based on the calculated feature amounts of the manuscript image. And executing a procedure of determining the document break of the document, and a procedure of separating the document image for each document according to the determined document break and storing the document image corresponding to each document as an individual electronic file. A computer-readable recording medium on which an image acquisition processing program to be recorded is recorded.